Think Like a Person Before Responding: A Multi-Faceted Evaluation of Persona-Guided LLMs for Countering Hate

TL;DR

This paper evaluates how different prompting strategies influence the quality, tone, and ethical safety of counter-narratives generated by large language models to combat online hate speech.

Contribution

It introduces a comprehensive framework for assessing LLM-generated counter-narratives across multiple dimensions including persona, readability, tone, and ethics.

Findings

Emotionally guided prompts produce more empathetic responses.

LLM-generated counter-narratives tend to be verbose and college-level in readability.

Safety and effectiveness concerns remain despite improved tone.

Abstract

Automated counter-narratives (CN) offer a promising strategy for mitigating online hate speech, yet concerns about their affective tone, accessibility, and ethical risks remain. We propose a framework for evaluating Large Language Model (LLM)-generated CNs across four dimensions: persona framing, verbosity and readability, affective tone, and ethical robustness. Using GPT-4o-Mini, Cohere's CommandR-7B, and Meta's LLaMA 3.1-70B, we assess three prompting strategies on the MT-Conan and HatEval datasets. Our findings reveal that LLM-generated CNs are often verbose and adapted for people with college-level literacy, limiting their accessibility. While emotionally guided prompts yield more empathetic and readable responses, there remain concerns surrounding safety and effectiveness.